Short, plain definitions of the 292 lean terms you hear most on the shop floor, from 5S and OEE to takt time and yokoten. Each entry answers the question first, then points to a longer guide on this site and, where there is one, the LeanSuite page that covers it. For what the letters in WCM, TPM, DMAIC or DOWNTIME stand for, see the lean acronyms list, and for the Japanese words and how to say them, the Japanese lean terms.
0–9
1-1-90 method
The 1-1-90 method is a way to roll out continuous improvement: start with 1 area and 1 pain point, run it for 90 days with a team, then expand what worked to the next areas and plants. It was named by Angad Singh and the LeanSuite team for the way they already work with manufacturers, and it is a general approach that needs no particular software. At day 90 the team decides whether to expand, run another round or stop.
The 5 Whys is a root cause analysis technique that asks “why?” repeatedly, with each answer becoming the next question, until the team reaches a cause it can act on. Five is a rule of thumb, not a fixed count. It works best on problems with a single chain of cause and effect, and each answer should be checked against facts at the gemba rather than guessed in a meeting room.
The 5G methodology is a problem investigation discipline that confirms five things before anyone proposes a cause: gemba (the actual place), gembutsu (the actual thing), genjitsu (the actual facts), genri (the principles that apply), and gensoku (the standard the process should follow). The first three are gathered by going to see; the last two compare what was found with how the process is supposed to work. It keeps root cause analysis grounded in evidence rather than opinion.
5S is a workplace organization method with five steps: Sort, Set in order, Shine, Standardize, and Sustain. Teams remove what is not needed, give everything a marked place, clean and inspect, write the standard down, and audit it so the area stays that way. A well-run 5S area makes abnormalities, missing tools, and safety hazards obvious at a glance.
The five steps, in order
1SortRemove what is not needed
2Set in orderGive everything a marked place
3ShineClean and inspect
4StandardizeWrite the standard down
5SustainAudit it so the area stays that way
1 Sort: remove what is not needed. 2 Set in order: give everything a marked place. 3 Shine: clean and inspect. 4 Standardize: write the standard down. 5 Sustain: audit it so the area stays that way.
A 5S audit is a regular, scored check of a work area against its 5S standard, using a fixed checklist for each of the five S’s. It is usually done by a supervisor or a rotating auditor on a set schedule, and the score is posted in the area so the team can see its trend. Findings become actions, which is how the Sustain step is kept up rather than left to good intentions.
5W1H is a questioning method that describes a problem or a plan by answering six questions: who, what, where, when, why, and how. In problem solving it produces a clear, fact-based problem statement before any cause analysis starts; in planning it makes sure an action has an owner, a scope, a place, and a date. Some versions add “how much” to make 5W2H.
The 6Ms are six categories used to sort the possible causes of a production problem: man (people), machine, method, material, measurement, and mother nature (environment). They are the usual bones of a fishbone diagram, and working through each one keeps a team from fixing only the cause it already suspects. The older five-category version, the 5Ms, leaves out mother nature.
The eight wastes of lean are the kinds of activity that use time and resources without adding value for the customer: defects, overproduction, waiting, non-utilized talent, transportation, inventory, motion, and extra processing. The first letters spell DOWNTIME, a common way to remember them. The original Toyota Production System named seven; unused talent was added later.
The eight wastes spell DOWNTIME
DDefects
OOverproduction
WWaiting
NNon-utilized talent
TTransportation
IInventory
MMotion
EExtra processing
D defects, O overproduction, W waiting, N non-utilized talent, T transportation, I inventory, M motion, E extra processing.
8D problem solving is a team method for fixing a serious or recurring problem in eight steps, called disciplines: form the team, describe the problem, contain it, find the root cause, choose permanent corrective actions, implement and verify them, prevent recurrence, and recognize the team. It is common in automotive and supplier quality, where customers often require an 8D report in response to a complaint.
A3 problem solving is a structured method that fits a whole problem on one A3-size sheet: background, current condition, goal, root cause analysis, countermeasures, and follow-up. The single page forces the team to agree on the facts and keep the story short enough for anyone to review. It follows the PDCA cycle and is often used by a leader to coach the person who owns the problem.
ABC analysis sorts inventory items by annual usage value, the annual quantity used times the unit cost, into three classes: A, the few items that make up most of the value; B, the middle group; and C, the many low-value items. A typical result is 10 to 20 percent of the items making up 70 to 80 percent of the value. The class sets how tightly each item is controlled, for example how often it is cycle counted and how tight its count tolerance is.
The acceptance quality limit (AQL), formerly called the acceptable quality level, is the worst average defect rate a sampling plan treats as acceptable over a continuing series of lots. It is one input to the ANSI/ASQ Z1.4 and ISO 2859-1 sampling tables, which use the lot size, the inspection level and the AQL to give the sample size and the accept and reject numbers. An AQL is not a promise that every lot is that good, since a lot with more defects can still pass on its sample, which is why critical characteristics are often checked on every part.
Acceptance sampling decides whether to accept or reject a lot by inspecting a random sample instead of every item. A single sampling plan is written (n, c): inspect n items and accept the lot if no more than c are defective. The plan's operating characteristic (OC) curve shows the chance of accepting a lot at each defect rate, which is where its producer's risk (a good lot rejected) and consumer's risk (a bad lot accepted) are read.
Andon is a visual, and often audible, signal an operator uses to call for help when a problem occurs at a station, such as a defect, a missing part, or a machine fault. Pulling the andon brings a team leader to the spot quickly, and on many lines the line stops if the problem is not cleared within a set time. It is a core tool of jidoka: stop and fix problems where they happen instead of passing them on.
What pulling the andon sets off
1
A problem at the station
A defect, a missing part or a machine fault.
2
The operator pulls the andon
A visual, and often audible, signal.
3
A team leader comes to the spot
Quickly, to stop and fix the problem where it happens.
4
The line stops if it is not cleared
On many lines, after a set time, so the defect is not passed on.
A problem at the station; the operator pulls the andon; a team leader comes to the spot quickly; on many lines the line stops if the problem is not cleared within a set time.
An annual objective is the year's share of a breakthrough objective: what the company or plant must achieve in the next twelve months to stay on the three to five year path, each measured by one KPI with a target and judged every month against it. On the X-matrix the annual objectives sit on the left, each linked to the breakthrough it serves and to the improvement priorities that will deliver it. In a cascade each facility writes its own version of the company's annual objectives and takes the share it can move. A plant that cannot say which annual objective a project serves usually has too many projects, not too few objectives.
At-risk behaviour, in David Marx's just culture model, is a choice whose risk the person did not see or mistakenly believed was justified, usually because the whole crew has drifted into the same shortcut and it has been paying off for months. It sits between human error, which is not chosen, and reckless behaviour, a conscious disregard of a substantial and unjustified risk. The response is to coach the person, remove whatever made the shortcut pay and fix the standard that could not be followed at line speed, because the same choice would be made by the next person on the next shift.
Attribute agreement analysis checks whether inspectors who sort parts into categories, such as accept or reject, agree with themselves, with each other and with a reference decision. Each inspector judges the same parts two or three times in a random order, and the results are reported as percent agreement, miss and false alarm rates, and often kappa. It is the measurement system analysis for visual and go or no-go inspection.
An attribute control chart tracks counts over time instead of measurements: the proportion or number of defective units (p and np charts) or the number of defects or defects per unit (c and u charts). Its control limits sit three standard deviations either side of the center line, using the binomial or Poisson model of the counts, and on the p and u charts they are worked out for each subgroup's sample size. A point beyond a limit or a long run on one side of the center line signals a special cause.
Autonomous maintenance is the TPM pillar in which machine operators take on basic equipment care: cleaning, inspecting, lubricating, tightening, and spotting abnormalities early. It is usually introduced in seven steps, starting with initial cleaning and moving toward operators managing their own equipment standards. Maintenance technicians are then freed for planned and more skilled work.
Balance efficiency is the total work content of a line divided by the number of operators times the longest station time. It shows how evenly the work is spread: for example, 252 seconds of work over 5 operators with a longest station of 52 seconds is 96.9%. Its complement, balance delay or balance loss, is the share of paid operator time spent waiting for the slowest station.
The balanced scorecard is a performance measurement framework introduced by Robert Kaplan and David Norton in the Harvard Business Review in 1992. It groups an organization's objectives and measures into four perspectives, financial, customer, internal process, and learning and growth, so financial results are read next to the capabilities that produce them. Kaplan and Norton later added the strategy map, which draws the cause-and-effect arrows between the objectives. A plant scorecard is usually a shorter, plant-level list of measures such as safety, quality, delivery, cost and people.
Batch production is making a set quantity of one product at a time, then changing the equipment over and making a set quantity of the next. It suits equipment shared by several products, but each batch waits as stock until it is used, so larger batches mean more inventory and longer lead times. In lean practice the batch size follows from the changeover time: shorter changeovers allow a shorter EPEI and smaller batches.
Behavior-based safety (BBS) is an approach to injury prevention that focuses on what people observably do at work. Trained observers watch tasks, record safe and at-risk behaviors, give immediate feedback, and use the pattern of observations to fix the conditions and systems that drive at-risk behavior. Done well, it is a conversation between peers, not a search for someone to blame.
A bottleneck is the step in a process with the least capacity relative to demand, so it limits the output of the whole line. Work piles up in front of it while the steps after it wait, which is the usual way to spot one on the floor. An hour lost at the bottleneck is an hour lost for the whole system, so, as the theory of constraints teaches, improvement there pays back most.
A boundary sample is a real part, or a photo or drawing when no part exists, that shows the limit of a visual defect: the worst condition an inspector still accepts. It is kept with a reject just beyond the limit, labelled, approved by quality and re-approved on a set date so it does not wear or fade. Boundary samples settle defects that are hard to put a number on, such as scratches seen at an angle, gloss or texture.
A bowling chart is the monthly scorecard of a hoshin kanri plan: one row per metric with the planned and actual value for each month, colored green when the actual meets the plan and red when it misses. Red months get a countermeasure from the metric's owner, and repeated reds usually get an A3. The name comes from its resemblance to a bowling score sheet.
Break-in work is unplanned maintenance done during a scheduled week that pushes a scheduled job out, such as a breakdown or an urgent repair that cannot wait for the next schedule. Logging each break-in with its hours and the job it displaced shows how much of the week reactive work takes and explains most of the gap in schedule compliance. Jobs pushed out go back to the backlog, usually first in line for the next week's schedule.
A breakdown analysis is a written investigation of one equipment breakdown: what was seen, how the machine was restarted, which part failed and why, the countermeasures, and a later check that the failure has not come back. In TPM practice it names which of the five causes of breakdowns was behind it and which standard changed as a result. Sites set their own trigger for one, such as a downtime threshold or any repeat failure.
A breakthrough objective is a three to five year goal that needs the business to change how it works, not just work harder: halving the plant's lead time, reaching zero recordable injuries, or doubling output from the same floor space. In hoshin kanri the breakthroughs sit at the bottom of the X-matrix and the rest of the matrix is read from them: the annual objectives are this year's step, the improvement priorities are the projects that take it, and the targets to improve are the KPIs that prove it. A company keeps three to five, because each one needs people and money for several years. If it could be reached by the normal improvement system, it is not a breakthrough.
Calibration compares a gauge or instrument with a reference standard of known value, records the difference, and adjusts the gauge if needed so its readings can be trusted. Each gauge is calibrated at a set interval against standards that are themselves traceable to national or international standards, and the readings are recorded as found, before any adjustment, and as left. When a gauge is found out of tolerance, the results it gave since its last good calibration have to be assessed, because parts may have been accepted or rejected wrongly.
Capital expenditure (CapEx) is money spent on assets that will be used for more than a year, such as machines, tooling and building changes, as opposed to operating expenses (OpEx) such as wages, materials and repairs, which are charged when they are spent. Capital is depreciated over the asset’s life and comes from a limited yearly budget, so most companies ask for a CapEx request with the options, costs, payback and NPV, approved at a level set by the amount. Where the line between capital and expense falls is set by each company’s accounting policy.
Cascading goals is the practice of translating a company goal into a goal for each level below it: company to facility, facility to department or team, department to person. Each level keeps the measure it can move and its share of the number, so the facility targets add up to the company target. Goal alignment is the result: every goal and initiative traces to a company goal, and the work that does not is visible.
Catchball is the back-and-forth dialogue used in hoshin kanri to agree goals and plans between levels of an organization. A leader proposes an objective, the team below responds with what it would take and what it can commit to, and the plan passes back and forth until both sides agree. The result is targets people understand and own, rather than numbers handed down.
A causal loss is the original problem, such as a machine breakdown or a changeover; a resultant loss is a cost it sets off elsewhere, such as operators waiting, overtime or scrap at the restart. Cost deployment in World Class Manufacturing separates the two so resultant costs can be traced back to the causal loss worth fixing. Each loss is counted once, and the causal losses are ranked by their own cost plus the resultant costs they cause.
Cellular manufacturing arranges the machines and workstations needed to make a family of similar products close together, often in a U-shape, so parts move from one step to the next with little transport or waiting. A cell is usually run by a small team that can move between stations as demand changes. It is the physical layout that makes one-piece flow practical.
Chaku-chaku, Japanese for “load-load,” is a way of running a cell in which the machines unload finished parts automatically and the operator’s only job at each station is to load the next part and move on. The operator walks the cell in process order, carrying one part at a time, so the cell runs as one-piece flow with minimal handling. It suits cells where machine cycles are automatic and short.
Change point management is logging every change to the people, machines, materials or methods of a process, the 4M, and checking the product more closely until the change is proven. Typical change points are a relief operator, a repair or new tool, a new material lot or supplier and a revised work instruction; some plants add Design and write 4M1D. Toyota-style plants call a change point henkaten and show current changes by station on a board.
A changeover matrix is a from-to table of changeover times between every pair of products a line runs, with the product running now in the rows and the next product in the columns. It shows that changeover time depends on the order: dropping an allergen or going from a dark product to a light one usually needs a fuller clean than the reverse. Planners use it to choose the run order with the least changeover time and to pick the biggest cells for SMED.
A check sheet is a simple form for recording how often something happens, and where or when, using tally marks made at the process as events occur. The categories are set in advance, for example defect types by hour or by machine, so the data is collected the same way every time and is ready to analyze. It is often the first of the seven basic quality tools a team uses, and its counts feed Pareto charts and histograms.
CILR stands for clean, inspect, lubricate, and retighten, the four basic equipment care tasks operators carry out as part of autonomous maintenance. Each machine gets a CILR standard that says which points to check, how, and how often, usually as a short checklist. Regular CILR catches loose fasteners, leaks, contamination, and wear before they turn into breakdowns.
Coaching kata is a short, fixed routine in which a coach asks a learner five questions about their improvement work: what is the target condition, what is the actual condition now, what obstacles are in the way and which one are you addressing, what is your next step and what do you expect, and when can we go and see what you learned. The learner follows the improvement kata, a repeating pattern of understanding the direction, grasping the current condition, setting the next target condition, and experimenting toward it. Practiced daily, the pair builds scientific problem-solving habits.
Compressed air leakage is the air a compressed air system loses through couplings, hoses, fittings, regulators, drains and valves, measured as a flow (cfm or L/s) or as a share of compressor output. The US Department of Energy says leaks often waste 20 to 30% of a compressor's output and that a leak program can hold them to 5 to 10%; the flow through a leak rises with pressure and with the square of the hole's diameter. Leaks blow whenever the system is pressurised, so their cost is the compressor power for that flow times the hours under pressure.
A computerized maintenance management system (CMMS) is software that keeps a plant’s maintenance records in one place: the asset list, work orders, preventive maintenance schedules, spare parts, and each machine’s repair history. It schedules recurring tasks, tracks each job from request to close, and turns the history into measures such as MTBF and MTTR. An enterprise asset management (EAM) system covers the same ground plus the wider asset life cycle, such as purchasing and replacement planning.
Condition-based maintenance (CBM) means checking equipment for a sign that a failure is coming and doing the repair only when the sign is found, instead of at a fixed age. The check can be an instrument reading such as vibration or oil analysis, or a simple inspection such as measuring wear depth. It works only when the sign appears long enough before the failure to act on, and reasonably consistently; reliability-centred maintenance tries it before any time-based task.
A connected worker platform is software that gives frontline workers the instructions, training, issue reporting and checklists they need on shift, on a phone, a tablet or a shared kiosk. It replaces paper binders, whiteboards and spreadsheets at the point of work, and sends what workers record to the people who act on it.
A containment action is the immediate step that stops a problem from reaching the customer or the next process while its root cause is found and fixed. Typical containment includes holding and sorting suspect stock, adding an inspection and posting a quality alert at the line. It protects the customer but does not remove the cause, so it is replaced by a permanent corrective action and then taken away.
Continuous improvement software is a system for managing improvement work from the first idea or reported problem to the measured result. It typically covers capturing ideas and issues from the floor, routing them for review and approval, running improvement projects such as kaizen events, A3s or PDCA cycles, and reporting participation, progress and results. It replaces suggestion boxes, spreadsheets and email trails with one record per improvement.
CAPA, short for corrective and preventive action, is the quality process for removing the causes of problems. Corrective action eliminates the cause of a nonconformity that has already happened so it does not recur; preventive action removes the cause of a potential problem before it occurs. A CAPA is closed only after its effectiveness has been verified.
Cost deployment is the World Class Manufacturing method that puts a money value on every loss in a plant, so improvement goes where the cost is highest. Losses are measured by process, split into causal losses and the resultant losses they set off, turned into money with agreed unit costs, and matched with the methods and projects that recover them. The work is usually laid out in matrices labelled A to E, from where losses occur to improvement cost against benefit.
Cost of poor quality (COPQ) is the cost a business carries because products or processes are not done right the first time. It covers internal failures such as scrap and rework and external failures such as returns, warranty claims, and complaints; broader definitions also count appraisal costs such as inspection. Putting a money value on defects helps rank which quality problems to fix first.
Cost per unit is what it costs to make one unit of a product: direct materials, direct labor and manufacturing overhead, added up and divided by the units made. Cost per good unit divides the cost of everything started by only the units that pass, so scrap is carried by the good ones. Overhead is charged to the product at a rate, such as budgeted overhead divided by budgeted direct labor hours, so part of the figure is an allocation rather than a measurement.
A cost-benefit analysis compares what a change will cost with what it will return, over the same years and in the same money terms, to decide whether it is worth doing and which of several options is best. Typical results are the net benefit per year, the payback period, the NPV and the benefit-cost ratio, the present value of the benefits divided by the present value of the costs, where a ratio above 1 means the benefits are larger. Hard benefits, where spending goes down, are usually judged first, with soft benefits shown separately.
Critical spare parts are the parts a plant keeps in stock because waiting for one would stop or endanger an asset that matters, and they could not be bought fast enough when it fails. Which parts count follows from the equipment criticality ranking, what each part's failure does, its lead time and whether it is single source or obsolete. For parts used a few times a year, the reorder point is often set with the Poisson model from usage and lead time.
A critical-to-quality (CTQ) characteristic is a measurable property of a product or process that customers care about, with a target and limits, such as a seal strength or an on-time delivery window. Six Sigma teams turn what customers say into CTQs in the Define phase, and a project's primary metric usually measures one of them.
Cross-training is the planned practice of teaching employees to perform jobs beyond their primary role. It gives a plant enough qualified people to cover absences, turnover, and swings in demand without stopping a line or depending on one expert. Progress is usually tracked in a skills matrix so gaps are visible before they cause a problem.
Cycle counting is counting a small part of the inventory on a regular schedule, so every item is counted over the year and the stock records stay accurate without shutting down for a full physical inventory. Each count is compared with the system quantity, and differences outside tolerance are recounted and traced to their cause, such as an unrecorded issue or a part in the wrong location. Most schedules use an ABC classification so that high-value items are counted most often.
Cycle time is how long one process step takes to complete one unit, as observed at the station. Comparing each station’s cycle time with takt time shows whether a line can keep up with customer demand and which station is the bottleneck. It is not the same as lead time, which is the total time a unit spends in the process, including waiting.
A daily huddle is a short stand-up meeting, usually 10 to 15 minutes at the start of a shift, where a team reviews yesterday's results against target, today's plan and open problems at its board. Problems get an owner and a date instead of being solved in the meeting, and what the team cannot fix is escalated. It is the front-line tier of tiered meetings.
Daily kaizen is the small, immediate improvement operators or supervisors make in their own area, with no approval process, done the same day or within the shift. Repositioning a tool, fixing a label or adding a hook are typical. Teams make it a habit with a short routine at their board every day, where each fix gets one owner and the fixes are counted every week.
The DART rate is the number of recordable cases with days away from work, restricted work or a job transfer (columns H and I of the OSHA 300 Log), multiplied by 200,000 and divided by the hours worked. Each case sits in one column only, the one for its most serious outcome, so a case with both days away and restricted days counts once. It shows how many recordable cases were serious enough to take a person off their normal job.
Defects per hundred units (DPHU) is the number of defects found per 100 units inspected: total defects divided by units inspected, times 100. Because one unit can carry several defects, it can be higher than the percentage of defective units, which counts each faulty unit once. Tracking both shows whether problems are spread across many parts or concentrated on a few.
A design FMEA (DFMEA) is an FMEA of a product design, done by the design team before the design is released, to find how the part could fail to do its job and fix the design while changes are still cheap. Each failure mode is rated for severity, occurrence and detection, where detection is how well design verification and testing would catch it, and the highest risks get actions. It assumes the part is made to the drawing; how the process could fail to make it to the drawing is the job of the process FMEA, which the DFMEA feeds through special characteristics.
Digital manufacturing readiness is how far a plant's work, systems and people are set up to run on digital information: work recorded where it happens, each number entered once, data reaching the people who act on it within the shift, and leaders and operators who use it. A readiness check scores process, technology and people together, because a tool the process and the people are not ready for adds little. Published frameworks include the Singapore Smart Industry Readiness Index and the acatech Industrie 4.0 Maturity Index.
Digital work instructions are step-by-step job instructions shown on a screen, tablet, or phone at the workstation instead of on paper. They usually combine short text with photos or video, and because they are updated in one place, every station sees the current version. They are how standard work and SOPs reach the operator on the shop floor.
DMAIC is the five-phase improvement method used in Six Sigma: Define, Measure, Analyze, Improve, and Control. The team defines the problem and goal, measures current performance, analyzes data to confirm root causes, improves the process, and puts controls in place to hold the gain. It suits complex, data-heavy problems where the cause is not obvious.
The five phases
1
Define
The problem and the goal.
2
Measure
Current performance.
3
Analyze
Data, to confirm root causes.
4
Improve
The process.
5
Control
Controls in place to hold the gain.
Define the problem and goal; Measure current performance; Analyze data to confirm root causes; Improve the process; Control, putting controls in place to hold the gain.
Dock-to-stock time is the time from a delivery being unloaded at the receiving dock to the parts being put away in their location and available to use. It covers checking the delivery against the purchase order, any inspection hold and putaway, so long times usually point to a queue at one of those steps. Material that sits at the dock cannot be used and is easily lost, which is why receiving logs record the putaway time.
Downtime reason codes are a fixed list of reasons an operator picks from when a machine or line stops, such as a changeover, a mechanical failure or waiting for material. Each code rolls up to one of the six big losses, so the minutes logged against it feed availability, performance or quality in OEE. A list that works is short, has one code for each kind of stop and is reviewed after its first month.
The eight pillars of TPM are the areas of work that total productive maintenance is organized into: autonomous maintenance, planned maintenance, quality maintenance, focused improvement, early equipment management, training and education, safety, health and environment, and office TPM. They rest on a foundation of 5S. Each pillar has its own owners and goals, but they depend on each other; autonomous maintenance, for example, relies on the training pillar.
Employee empowerment is giving people the authority, information and resources to make decisions about their own work without asking a manager each time. In practice it means decisions agreed in advance, with clear limits and a named person to consult or escalate to, plus the training to use them. The empowerment guide on this site calls it the opposite of micromanagement.
Employee engagement is how committed people are to their work and to the organization's goals, shown in the effort they give beyond the minimum. It is not the same as satisfaction or happiness. The engagement guide on this site names six drivers: recognition, feedback, culture, team, empowerment and development.
An employee recognition program is the set of rules a workplace uses to notice and thank people for specific actions: what is recognized, who can nominate, which rewards are given and who approves them. In a plant it often covers safety reports, improvement ideas, quality catches and helping other teams, and a regular check shows whether recognition reaches every team and shift, not only the people managers see most.
An energy performance indicator (EnPI) is a measure of a site's energy performance that can be compared over time, such as energy per unit made or the ratio of actual energy to what a baseline model expected. The energy baseline is the reference period or model it is compared with. A baseline that accounts for what moves energy use, such as output and weather, stops a slow month or a mild winter from looking like savings. ISO 50001 uses both terms, and the US DOE 50001 Ready protocol uses 12 months for the baseline.
An engineering change request (ECR) is the formal request to change a part, drawing, process or document. It says what changes and why, lists everything the change touches, such as tooling, stock on hand, work instructions, the PFMEA and customer approvals, and is reviewed and approved before anything changes on the floor. Once approved, an engineering change notice (ECN) tells everyone affected what changes and when it takes effect; some companies call this step an engineering change order (ECO).
EPEI is how often each product or part number comes round in a production sequence. If a machine makes every part assigned to it at least once every three days, its EPEI is three days. Shorter is better because it means smaller lots and less stock, and it depends mainly on changeover times and the number of parts the machine runs.
Equipment criticality analysis ranks a plant's assets by the consequences of their failure, usually by scoring safety and environment, quality, production impact, failure frequency, repair time and spare parts lead time. The weighted total sorts assets into classes, often A, B and C, that decide which get condition monitoring and critical spares, which get time-based preventive maintenance and which can run to failure. Criteria, weights and cut-offs are set by each site, and many add a rule that a top safety score makes an asset class A.
Equivalent annual cost (EAC) is the even yearly amount that has the same present value as all of an asset's costs over its life: purchase and installation, yearly operating and maintenance costs, less what it is worth at the end. It is the present value of those costs times the capital recovery factor r ÷ (1 − (1 + r)^−n), at discount rate r over n years. Because it is a cost per year, it compares a machine that will last 4 more years with one that will last 10, which is how repair or replace decisions are made; the number of years with the lowest EAC is the asset's economic life.
An escalation matrix is a table that says who is called when a problem cannot be solved where it starts, and when: for each type of issue, who responds first and how fast, and who is called next if the problem is not contained within a set time. It is usually set out by issue type and by tier, from the team leader up to plant leadership. It sits behind andon calls, issue tags and tiered meetings, so each problem reaches someone who can fix it in time.
An F-tag is a tag hung on equipment to mark an abnormality found during inspection or daily work, such as a leak, a loose bolt, contamination, or a damaged guard. The F stands for fuguai, Japanese for an abnormality or malfunction. Each tag is logged and stays visible on the machine until the problem is fixed, and tags are often split by whether operators or maintenance should resolve them.
Failure mode and effects analysis (FMEA) is a structured way to find how a product or process could fail before it does. For each failure mode the team rates severity, occurrence, and detection, then acts on the highest risks first, traditionally ranked by a risk priority number (RPN) and in newer methods by action priority. Process FMEAs are a standard part of launching and controlling manufacturing processes.
A failure-finding task is a test or inspection done at a set interval to find out whether a hidden function has failed, such as lifting the float of a level switch to see that the pump stops. It is used for protective devices that give no sign when they fail, when no preventive task works. The interval is chosen so the device is unlikely to be failed on the day the thing it protects needs it.
Fault tree analysis (FTA) is a top-down root cause method that starts from one undesired event, such as a machine failure or a safety incident, and works backward through the combinations of events that could cause it. The causes are drawn as a tree joined by AND and OR logic gates, which shows whether one failure alone or several together are needed to produce the top event. It suits complex equipment and safety problems where several causes interact.
A FIFO lane is a capped queue between two processes in which parts leave in the same order they arrive. When the lane holds its maximum, the process feeding it must stop until the process after it takes a unit out, so the cap itself is the pull signal and no kanban card is needed. It fits two adjacent processes with one customer for the lane; a supermarket fits better when lot sizes differ, flow splits or merges, or the processes are far apart.
A fire extinguisher inspection is the scheduled check that each portable extinguisher is in its place, charged and ready for use. In the United States, OSHA 29 CFR 1910.157 requires a visual inspection every month, an annual maintenance check with the date recorded, and a hydrostatic test at 5 or 12 year intervals depending on the type. The employer's written procedure, local law, NFPA 10 and the insurer can require more.
First pass yield (FPY) is the share of units that come through a process step right the first time, with no rework, repair, or scrap. It is calculated as good units made first time ÷ units started × 100. Because reworked units are left out of the good count, FPY exposes the rework that a final yield figure can hide.
A first piece inspection checks the first part made after a setup, a change or a restart against the drawing or the control plan before the run continues. It confirms the process is making good parts now, which is why it follows changeovers, repairs and other change points. It is not the same as a first article inspection, which verifies a new part or a design change once, in full.
A fishbone diagram, also called an Ishikawa or cause-and-effect diagram, is a visual tool for brainstorming the possible causes of a problem. The problem sits at the head of the fish and causes are grouped along the bones, often under categories such as man, machine, method, material, measurement, and environment. It organizes ideas for investigation; the likely causes still have to be verified with data.
The five lean principles describe how to remove waste from a process: specify value from the customer’s point of view, identify the value stream, make the value-creating steps flow, let the customer pull, and pursue perfection. They were set out by James Womack and Daniel Jones in the book Lean Thinking. The cycle repeats, because each pass exposes more waste to remove.
A fixed-position stop is a way of stopping an assembly line at the end of the work cycle, at a set position, when a problem called on the andon cannot be solved within that cycle. The team leader has until the line reaches that position to fix the problem, so many calls can be cleared without stopping the line. It is one way of putting jidoka into practice on a paced line.
Floor marking is the use of painted or taped lines, colors and symbols on the floor to show aisles, work areas, storage and material locations, and hazards, so anyone can see where things belong and where people and trucks should go. A common scheme uses yellow for aisles and work cells, red for defects and scrap, and striped tape for areas to keep clear, but there is no single universal standard. OSHA requires permanent aisles to be marked and uses red for fire equipment and yellow for physical hazards, and leaves the rest of the scheme to each site, which should write it down and use it the same way in every area.
Food GMP, short for current good manufacturing practice, is the set of baseline conditions a food plant keeps in place for every product: personnel hygiene, plant and grounds, cleaning and pest control, water and hand-washing facilities, equipment, process controls and storage. In the US it is set for human food by 21 CFR Part 117, subpart B; the Codex General Principles of Food Hygiene (CXC 1-1969) cover the same ground internationally. HACCP plans and food safety plans are built on top of it.
A Gantt chart is a project schedule drawn as horizontal bars on a time scale, one bar per task from its start to its finish, with milestones marked as points. Links between tasks show what each one waits for, so a late task shows what it holds up. Improvement teams use one to schedule a project after the charter and to compare planned with actual progress at each review.
Gauge R&R (repeatability and reproducibility) is a study that measures how much of the variation in measurement results comes from the measurement system itself. Repeatability is the variation when one person measures the same part several times with the same gauge; reproducibility is the variation between different people measuring the same parts. The combined result is compared with the part tolerance or the process variation to decide whether the gauge is fit for use.
Gemba is a Japanese word meaning “the actual place,” and in lean it means where value is created, most often the shop floor. The principle is that problems are best understood by going there to see the process, the product, and the people directly, rather than relying on reports.
A gemba walk is a leader’s planned visit to the place where work is done to observe the process, ask questions, and learn from the people doing the work. The aim is to see real conditions and support problem solving, not to inspect people or hand out orders on the spot. What the leader sees is noted and followed up as actions.
Genchi genbutsu, Japanese for “actual place, actual thing,” is the Toyota principle of going to the source to see a problem for yourself before deciding what to do. Leaders and engineers check the real process, parts, and data at the gemba instead of relying on reports or assumptions. It is the thinking behind gemba walks and the 5G methodology.
Goal alignment is the state in which every initiative, project and measure in a plant can be traced to a company goal, and every company goal has work underneath it. Cascading is the top-down handoff; alignment is the result when goals at each level are visible and linked, including goals a team proposed from the bottom up. The simplest check is a list of every running project and kaizen with the goal each one serves: work that serves none goes in an Other column, and the review asks whether it should continue. Kaplan and Norton reported in Harvard Business Review (2005) that on average 95% of a company's employees were unaware of, or did not understand, its strategy, which is why alignment has to be visible on the floor and not only in the plan.
Goal health is a one-page view of every company goal against every facility for the last closed month, each cell marked Met or Missed from that facility's KPI, so leadership sees in one look which goals are on track everywhere, which are missed at one site and which are missed across the company. The counts worth reading off it are goals missed this month, goals missed two or more months in a row, and goals with no KPI attached, which cannot be judged at all. A bowling chart shows one site's goals across the months; goal health shows every site for one month, and the two together are the agenda for the monthly review. A goal missed two months running gets an A3 or a project, not another month.
HACCP (hazard analysis and critical control points)
HACCP, short for hazard analysis and critical control points, is a food safety system that finds the biological, chemical and physical hazards in a product and its process and controls the significant ones at critical control points (CCPs), each with a measurable critical limit, monitoring, corrective action, verification and records. It follows the seven principles in the Codex Alimentarius General Principles of Food Hygiene (CXC 1-1969) and rests on prerequisite programmes, such as cleaning and hygiene, that must already work.
Hansei is the Japanese practice of honest self-reflection: looking back at what went wrong or could have gone better, admitting one’s own part in it, and committing to a concrete change. At Toyota it is expected even after a success, on the view that there is always something to improve. Hansei meetings, or hansei-kai, are held at the end of a project or event to capture the lessons before they are lost.
Hard savings are reductions in spending that show up in the accounts, such as less material bought, fewer overtime hours paid or a contract cancelled. Soft savings are costs avoided or time and space freed without anyone paying less yet, such as operator hours freed with no change in headcount or overtime. Most CI programs report the two separately and have Finance verify the hard savings before they count toward a target, so the CI team and the plant controller quote the same number.
Heijunka, or production leveling, is the practice of smoothing the volume and mix of production over a period instead of following the ups and downs of incoming orders. Rather than building all of product A on Monday and all of product B on Tuesday, a leveled schedule makes some of each every day in a repeating sequence. It reduces mura (unevenness), keeps upstream processes and suppliers at a steady pace, and makes just-in-time workable.
Batch by day against a levelled sequence
All of A on Monday, all of B on Tuesday
A
A
A
A
A
B
B
B
B
B
Some of each every day, in a repeating sequence
A
B
A
B
A
B
A
B
A
B
Unlevelled: all of product A on Monday and all of product B on Tuesday. Levelled: some of each every day in a repeating sequence.
The hierarchy of controls ranks ways to protect workers from a hazard, from most to least effective: elimination, substitution, engineering controls, administrative controls and personal protective equipment. NIOSH notes that the top three work without relying on people, while the bottom two need constant effort from workers and supervisors. Incident investigations and risk assessments use it to choose the strongest action that can be put in place.
A high reliability organization is one that does hazardous, complex work for long periods without a serious accident, such as an aircraft carrier flight deck, air traffic control or a nuclear power plant. It does not avoid every failure; it notices small ones early and stops them growing. Karl Weick and Kathleen Sutcliffe describe five principles behind it in Managing the Unexpected (2001): preoccupation with failure, reluctance to simplify, sensitivity to operations, commitment to resilience and deference to expertise.
A histogram is a bar chart that shows how many measurements fall into each of a series of equal ranges, so the shape of a process’s variation can be seen at a glance. It shows whether results are centered on the target, how widely they spread, and whether the pattern is skewed or has two peaks, which can mean two machines, shifts, or material lots are mixed together. Drawing the specification limits on it gives a quick visual check of process capability.
Hoshin kanri, also called policy or strategy deployment, is a method for linking a company’s few breakthrough goals to the work of every level of the organization. Goals are cascaded through a back-and-forth negotiation called catchball, often recorded on an X-matrix, and progress is reviewed on a regular PDCA cycle. It keeps improvement effort focused on what matters most to the business.
An hour-by-hour board is a visual board at a line or cell that shows planned output against actual output for each hour of the shift, with a short note of the reason whenever an hour falls short. Because it is updated every hour, a missed target becomes a problem the team can act on during the same shift instead of something found in an end-of-day report. The recorded reasons feed the daily meeting and show which losses to attack first.
An improvement priority is the project, kaizen or programme that moves an annual objective: the work, as opposed to the goal. In hoshin kanri the priorities sit at the top of the X-matrix, each marked against the objectives it supports and the KPIs it should move, with a lead owner and the people who support it in the resources column. A plant chooses few enough that each has an owner with the hours to do it, usually one to three per objective, and a priority that supports no objective is cut. Each priority is delivered through the normal improvement system, kaizen events, projects and A3s, so that strategy and daily improvement are the same work and not two lists.
Incoming inspection, also called receiving inspection, is checking purchased material when it arrives, before it is used or stocked: the paperwork and certificates, the packaging and condition, and a sample checked against the drawing. How many parts to check comes from a sampling plan, often from the ANSI/ASQ Z1.4 or ISO 2859-1 tables with an inspection level and AQL the site chooses. A lot that fails is held and recorded on an NCR, and a SCAR goes to the supplier when the cause is theirs.
An individuals and moving range (I-MR) chart is the control chart for one reading at a time, such as one viscosity per batch or one number per day. The individuals chart plots each reading against limits of the mean plus and minus 2.66 times the average moving range, and the moving range chart plots the difference between each reading and the one before, with an upper limit of 3.267 times the average moving range. It is used when readings cannot be grouped into subgroups, where an X-bar and R chart would be used instead.
An internal audit is a planned check, by the organization's own trained auditors, that its quality management system works as intended and meets its requirements: the site's procedures, the customer's requirements and a standard such as ISO 9001. Auditors who do not work in the area sample records, watch the work and talk to the people doing it, then grade each finding against a stated requirement. Findings go to corrective action and are closed only after someone checks that the action worked.
The internal rate of return (IRR) is the discount rate at which a project’s net present value is exactly zero, which makes it the yearly return the project earns on its investment. For a normal project, with the money spent first and the returns after, an IRR above the hurdle rate set by Finance means a positive NPV at that rate. IRR is found by trial or with a spreadsheet function; when comparing options of different sizes, NPV is the safer guide, because a small project can have a high IRR and still add less value.
Inventory record accuracy is the share of counted items whose physical count matches the system record within an agreed tolerance: counts within tolerance divided by counts checked, times 100. Tolerances are usually tighter for high-value A items than for C items. Low accuracy shows up as shortages of parts the system says are in stock, extra safety stock, and people keeping their own counts.
Inventory turnover is how many times a business uses up and replaces its inventory in a period: cost of goods sold divided by the average inventory value at cost. Days on hand is the days in the period divided by turnover, which is how long the stock would last at the current rate of use. A higher turnover means less money tied up in stock; a turnover that is too high can mean shortages and expediting.
Is/Is Not analysis is a problem definition method that compares what, where, when, and how much a problem occurs with where and when it could occur but does not. The differences between the two columns point to what is distinctive about the problem and rule out causes that cannot explain the pattern. It is used early in 8D and other root cause methods to narrow the search before testing theories.
ISO 45001 is the international standard for occupational health and safety management systems, first published in 2018. It sets requirements for how a site finds hazards, controls risks using the hierarchy of controls, involves workers, meets its legal requirements, audits itself and corrects problems. Certification to it is voluntary and is not the same as complying with safety law such as OSHA rules.
Jidoka, often translated as autonomation or automation with a human touch, is the lean principle of building quality in by stopping the process as soon as an abnormality occurs. Machines detect problems and stop themselves, and people are empowered to stop the line, so defects are not passed downstream. It is one of the two pillars of the Toyota Production System, alongside just-in-time.
Jishuken is a hands-on improvement workshop in which a group of managers or engineers studies a real process at the gemba and improves it there, usually over a few days, to build their own kaizen skills. The Japanese term roughly means “self-directed study.” It developed at Toyota, which uses it to train leaders and to support improvement in its own plants and at suppliers, with the lessons spread through yokoten.
A job breakdown sheet is the instructor's teaching notes in TWI Job Instruction: a job split into its important steps, the key points within each step, and the reason for each key point. Key points are anything that could make or break the job, injure the worker, or make the work easier. It is kept short so the trainer can read it at a glance while teaching.
A job cover matrix is a chart that shows, for each job or station, which people are qualified to cover it and how many backups exist. It exposes single points of failure, such as a machine only one person can run, so supervisors can plan cross-training and staff each shift with confidence. It is closely related to the skills matrix but organized around covering roles.
A just culture is a way of responding to safety events that separates what a person did into honest error, at-risk behaviour and reckless conduct, and answers each differently: fix the system and support the person after an error, coach and remove the cause of drift after at-risk behaviour, and reserve discipline for a conscious disregard of a substantial and unjustified risk. It is not a no-blame culture. Its purpose is to keep people reporting, because blame for honest error silences the near-miss reports a plant learns from.
Just-in-time (JIT) is a production approach in which each process makes and delivers only what the next process needs, when it needs it, and in the quantity it needs. It relies on pull signals such as kanban, level scheduling, and short changeovers, and it keeps inventory low so problems surface instead of being hidden by stock. JIT and jidoka are the two pillars of the Toyota Production System.
Kaikaku, Japanese for “reform” or “radical change,” is a large, fast step change to a process or system, such as moving a line into cells or redesigning a whole value stream. Where kaizen improves in small, continuous steps, kaikaku is usually led from the top and changes how the work is organized. Plants follow a kaikaku with kaizen to stabilize and refine the new way of working.
Kaizen is the practice of continuous improvement through small, frequent changes made by the people who do the work. The Japanese word means “change for the better.” In a plant it shows up as daily problem solving, employee suggestions, and focused improvement events, all aimed at reducing waste and making work easier, safer, and more reliable.
A kaizen burst is the starburst symbol on a value stream map that marks a change needed to reach the future state, such as a shorter changeover or a new supermarket. On its own it is only an idea. In a value stream plan each burst becomes a project with one owner, a target, dates and a way to check that it worked.
A kaizen event, also called a kaizen blitz or rapid improvement event, is a short, focused project, usually lasting a few days to a week, in which a cross-functional team improves one defined process. The team maps the current state, finds the waste, tests changes on the floor, and standardizes the new method before the event closes. Open actions are tracked afterward so the gain holds.
A kaizen newspaper is the action list a kaizen team keeps for the items it cannot finish during the event: the problem, the action, who owns it and when it is due, with a status mark updated until each item is closed. It is usually posted in the area and reviewed at the 30, 60 and 90 day follow-ups, so the event's open work stays visible after the team disbands.
Kaizen teian is a system for collecting and acting on small improvement ideas from individual employees. Most ideas are changes people can make in their own area, and the emphasis is on quick review, recognition, and implementation rather than large savings per idea. It is the individual, everyday side of kaizen, alongside team-based kaizen events.
A kamishibai board is a board of double-sided cards, one per check, used to schedule and track short audits of standard work and process conditions. A card shows green when its check has been done and the process met the standard, and red when the check found a problem or has not been done, so anyone can see audit status at a glance. It is a simple visual way to run leader audits such as layered process audits.
Kanban is a signal, traditionally a card, that authorizes a process to produce or move a set quantity of parts. When a downstream process uses material, the kanban goes back upstream as the instruction to replace it, so production is pulled by actual use instead of pushed by a forecast. The number of kanban in the loop caps how much inventory can build up.
The kanban loop
1
Downstream uses material
A set quantity of parts is consumed.
2
The kanban goes back upstream
The card is the instruction to replace it.
3
Upstream produces or moves the set quantity
The number of kanban in the loop caps the inventory.
then the next kanban, pulled by actual use, not a forecast
A downstream process uses material; the kanban goes back upstream as the instruction to replace it; the upstream process produces or moves the set quantity; and the loop repeats, pulled by actual use. The number of kanban in the loop caps how much inventory can build up.
A key result is a measurable outcome that shows an objective has been reached: a number, where it starts, where it must get to and by when, such as cutting the average changeover on the filling lines from 42 to 25 minutes by the end of June. Each objective has three to five key results, and the test of a good one is that it measures a result, not an activity: training 40 operators on SMED is a task, every changeover under 25 minutes is a result. In a plant most key results are KPIs the team already sees on its board, which keeps the goal and the daily measure the same number. Key results are graded at the end of the cycle, and a miss is read as information about the plan, not as a mark against the owner.
A manufacturing KPI dashboard shows a small set of key performance indicators for a plant or area against targets, often grouped as safety, quality, delivery, cost and people, with a color when a number is off target and a trend over recent weeks. Each indicator needs one written definition and one data source, so every shift reads it the same way. Its job is to show where to act this week, not to report numbers.
A KPI tree is a diagram that puts one top-level measure at the root and splits it, level by level, into the smaller measures that produce it, joined by arithmetic: parts that add up, or rates that multiply. OEE is the plant standard, availability times performance times quality, with the minutes lost to breakdowns, changeovers and small stops underneath. The splitting stops at measures one team owns and can change, so a red top number can be traced to the branch that moved it and to the person who acts. It explains why a number moved; a bowling chart or a KPI dashboard shows whether it did.
A layered process audit (LPA) is a short, frequent check that key process steps and controls are being followed, carried out by several layers of the organization, from supervisors up to plant leadership. Each layer audits on a set schedule from a standard question list, and any nonconformance is corrected on the spot or assigned as an action. LPAs are widely used in automotive manufacturing to confirm that high-risk controls stay in place.
Lead time is the total time from the start of a process to its end as the customer or the next stage experiences it, for example from order to delivery. It includes processing, waiting, inspection, and transport, so it is usually much longer than the time spent actually working on the product. Common types are customer lead time, material lead time, production lead time, and cumulative lead time.
Leader standard work is a written routine of the daily, weekly, and monthly tasks a leader performs to sustain the operating system, such as gemba walks, tiered meetings, audits, and checking visual boards. Writing it down makes the leader’s own work visible and repeatable and protects time for coaching and problem solving. It is the management counterpart to operator standard work.
Lagging indicators measure outcomes after they have happened, such as recordable injuries, lost-time incidents, or customer returns. Leading indicators measure the activities and conditions that come before those outcomes, such as near-miss reports, safety observations, audits completed, or hazards closed. Tracking both lets a plant act on risks before they show up in the results.
Lean manufacturing is a way of running production that aims to deliver what the customer values with the least waste of time, material, effort, and money. It grew out of the Toyota Production System and combines principles such as flow, pull, and continuous improvement with tools such as 5S, standard work, and value stream mapping. In practice it depends as much on how people solve problems every day as on the tools.
A lean readiness assessment checks, before a lean rollout, whether the organization has the foundation lean depends on: leadership commitment, a culture that surfaces problems, processes stable enough to standardize, and people with the time and experience to take part. Several people score the same statements on their own, and the weakest factor and the gaps between their views show what to fix first and where to pilot. Repeated later, it shows whether the gaps have closed.
A lean transformation is an organization-wide change to working by lean principles, covering how work flows, how problems are surfaced and solved, and how leaders manage, rather than applying a few tools in one area. It usually starts in a pilot area or value stream and spreads as people and systems are ready. Because it changes daily habits and management routines, it is treated as an ongoing change, not a project with an end date.
Line balancing is the practice of distributing work across the stations of a line so that each station’s workload is close to takt time and to the others. A balanced line has less waiting and fewer bottlenecks, and it needs fewer operators to meet the same demand. Yamazumi charts are the usual tool for seeing and redistributing the work.
Little’s Law says that in a stable process the average time a unit spends in the process equals the average number of units in it divided by the rate at which they leave: lead time = WIP ÷ throughput. 1,600 parts waiting in front of a step that uses 400 a day means about 4 days of waiting. It is how value stream maps turn counted inventory into days.
Lockout/tagout (LOTO) is the safety procedure that isolates a machine from all hazardous energy sources and keeps it isolated while people service or maintain it. Each authorized worker applies their own lock and tag to the energy-isolating devices, and the absence of energy is verified before work begins. In the United States, the requirements are set by OSHA standard 29 CFR 1910.147.
Machine guarding is the use of guards and safeguarding devices to keep people away from a machine's point of operation, nip points, rotating parts, power transmission parts and flying chips. A guard is a physical barrier fixed to the machine; a device such as a light curtain or two-hand control detects or blocks access. In the United States the general rule is OSHA 29 CFR 1910.212, with 1910.219 for belts, gears and shafts.
A machine hour rate is the cost of running one machine for one productive hour: its yearly costs (depreciation, floor space, insurance, maintenance, tooling and, under some conventions, interest on the capital) divided by the hours it actually runs, plus the energy it uses per running hour. Because the yearly costs are spread over the hours run, a machine that runs fewer hours has a higher rate. Operator labour is usually kept as a separate rate and added for quotes.
Maintenance backlog is the approved maintenance work that has not been done yet, usually counted in weeks: estimated hours of backlog ÷ crew hours available per week. 360 hours of backlog with 120 crew hours a week is 3 weeks. A backlog that keeps growing means work arrives faster than the crew can do it; one near zero can mean work is not being planned ahead.
A maintenance job plan is the package prepared for one job before it is scheduled: the steps in order, the hours for each craft, the parts checked against stock, special tools, lockout and permits, and the drawings. A job is ready for the weekly schedule when everything in the plan is in place.
A master sample is a part or test piece in a known condition that is used to prove a gauge, sensor or error-proofing device still works. A known-bad sample carries the exact defect the device must catch, so a device that lets it through has failed. Master samples are labelled, kept in one place and controlled so they do not change, and they are never used as production parts.
A master sanitation schedule is a food plant's standing list of every cleaning task: the area and equipment, what is done, how often, the method and chemical, who does it and when it is next due. A pre-operational (pre-op) verification is the check before start-up, by someone other than the cleaner, that cleaned equipment is clean, usually by sight and sometimes with an ATP swab against a limit the site sets. Failed equipment is recleaned and rechecked before release.
Match marks are paint lines drawn across a nut or bolt head and onto the surface beside it once the fastener has been tightened to its torque. If the fastener turns, the line breaks or no longer lines up, so anyone walking past can see it has loosened without putting a wrench on it. They are one of the most common visual controls in autonomous maintenance; a broken line is tagged, not just retightened.
Mean time between failures (MTBF) is the average time a repairable machine runs between breakdowns. It is calculated as total operating time ÷ number of failures in that period, so a machine that ran 1,000 hours with 5 failures has an MTBF of 200 hours. A rising MTBF shows equipment is becoming more reliable; it is usually tracked alongside mean time to repair (MTTR), the average time to restore a machine after a failure.
Mean time to repair (MTTR) is the average time it takes to restore a failed machine to working order. It is calculated as total repair time ÷ number of repairs in that period, so 20 hours spent on 5 repairs gives an MTTR of 4 hours. Plants differ on whether repair time covers the whole outage, including diagnosis and waiting for parts, or only hands-on work, so the definition should be stated with the number; a falling MTTR means breakdowns are being fixed faster.
Measurement system analysis (MSA) is a set of studies that checks whether a measurement system, meaning the gauge, the method, the people, and the conditions, gives data good enough to base decisions on. It looks at bias, linearity, stability, repeatability, and reproducibility, with gauge R&R the most common study. If the measurement system is not adequate, control charts, capability studies, and pass/fail decisions built on its data cannot be trusted.
Measurement uncertainty says how far a measured value could reasonably be from the true value of what was measured. Under the GUM method each source, such as repeatability, the reference standard's calibration, resolution and temperature, is turned into a standard uncertainty and the sources are combined by root sum of squares. That combined value is multiplied by a coverage factor k, usually about 2, to give the expanded uncertainty U, and the result is reported as y ± U.
Micro-stops are brief stoppages, from a few seconds to a few minutes, that interrupt a machine or line without being logged as downtime, such as jams, misfeeds, or sensor trips an operator clears by hand. Because each one is short they are rarely recorded, yet together they can be a large hidden loss. In OEE they count as a performance loss and are one of the six big losses.
A milk run is a fixed route run at fixed times that delivers or collects material at several stops in one loop, instead of a separate trip for each. Inside a plant it feeds lines from a supermarket, often run by a water spider; outside, a truck collects small loads from several suppliers. Fixed timing lets each stop plan around the next visit and keeps the stock between visits small.
A minimal cut set is the smallest combination of basic events in a fault tree that together cause the top event: take away any one of them and that combination no longer causes it. A cut set with a single event is a single point of failure, and a cut set with two or more events shows where independent safeguards protect the system. With independent events, the probability of a cut set is the product of its events' probabilities.
A mizusumashi, or water spider, is a material handler who runs a fixed, timed route delivering parts to production lines and collecting empty containers and kanban. By keeping materials at the point of use, the role lets operators stay at their stations doing value-added work. Because the route and timing are standardized, the water spider also carries the pull signals between supermarkets and the line.
A mock recall is a practice run of a product recall: someone names a finished lot, and the team traces it one step back to every ingredient and packaging lot and its supplier and one step forward to every first customer, reconciles the quantities and times how long it took. Food safety schemes such as SQF and BRCGS require a test at least once a year. The gaps it finds get an owner, a fix and a re-test.
A model line is one production line or area where a plant first puts lean practices in place and proves they work before spreading them to other areas. The team builds the foundations there, such as 5S, standard work, visual management and a daily team meeting, then flow and pull, and writes what works into standards the next areas start from. Concentrating the effort on one line builds people's skill and gives the rest of the plant a working example to copy.
A monthly operations review is the leadership meeting that checks the year's goals against their KPIs once the month closes, in a fixed order, usually safety, quality, delivery, cost and people, with each goal's owner saying whether it was Met or Missed, the cause of any miss and the countermeasure with a name and a date. The bowling chart or goal health grid is the agenda, so nobody builds slides; owners prepare their part before the meeting, and reds get the time while greens get a sentence. It is the Check of hoshin kanri's PDCA cycle, between the daily tier meetings and the quarterly step back where the plan itself can change. The output is a short list of decisions and actions, reviewed first at the next meeting.
Muda, mura, and muri are the three types of waste in the Toyota Production System. Muda is any activity that adds no value for the customer, mura is unevenness in workload or flow, and muri is overburden on people or equipment. They are linked: uneven schedules (mura) push people and machines past their limits (muri), which creates waste (muda).
How the three link
Mura
Uneven schedules
Unevenness in workload or flow.
Muri
Push people and machines past their limits
Overburden on people or equipment.
Muda
Which creates waste
Activity that adds no value for the customer.
Uneven schedules (mura) push people and machines past their limits (muri), which creates waste (muda).
A near miss is an unplanned event that could have caused injury, illness, or damage but did not, often by luck or a last-second reaction. Reporting and investigating near misses lets a plant fix hazards before someone gets hurt. A strong reporting culture depends on reports being easy to make, free of blame, and visibly acted on.
Nemawashi is the practice of building consensus for a proposal informally, one to one with each person affected, before it goes to a formal decision. The word comes from Japanese gardening, where it means preparing the roots before moving a tree. It surfaces concerns early, improves the proposal, and makes the decision faster and easier to carry out.
Net present value (NPV) is the value today of a project’s future cash flows minus the investment. Each year’s net cash flow is divided by (1 + r) raised to the number of years from now, where r is the discount rate set by Finance, the results are added up, and the investment is subtracted. A positive NPV means the project returns more than the rate Finance requires; unlike simple payback, NPV counts every year of the project and the time value of money.
A no-blame culture is one in which nobody is held to account for any safety event, however it came about. It is often confused with a just culture, which refuses to blame honest error and drift but still answers reckless conduct with discipline. James Reason argued in 1997 that a no-blame culture is neither feasible nor desirable: a small share of unsafe acts are deliberate and deserve a sanction, and a blanket amnesty loses the trust of the people who follow the rules, so the plant ends up with less reporting rather than more.
Noise dose is a worker's daily noise exposure as a percentage of the amount allowed: for each period at one level, the time spent divided by the time allowed at that level, added up and multiplied by 100. Under OSHA 29 CFR 1910.95 the time allowed halves for every 5 dB above 90 dBA, so 100% is the permissible exposure limit and 50%, an 8-hour time-weighted average (TWA) of 85 dBA, requires a hearing conservation program. NIOSH halves the time every 3 dB above 85 dBA, so the same noise gives a higher NIOSH dose.
A non-conformance report (NCR) is a record that a product, material, or process did not meet a specified requirement. It documents what was found, where, and how much is affected, along with the decision on what to do with it, such as rework, scrap, use as is, or return to the supplier. Serious or recurring nonconformances are escalated to a CAPA to find and remove the root cause.
An obeya, Japanese for “big room,” is a dedicated room where a cross-functional team keeps all the key information for a project or strategy on the walls, such as goals, schedules, metrics, problems, and countermeasures. The team meets there regularly, standing at the visuals to review status and make decisions quickly. It began at Toyota in product development and is now also used to manage plant performance and strategy.
An objective is the qualitative half of a goal: a short statement of the change a company, plant or team wants to make, such as making the packing lines the most reliable in the plant, written so people remember it and can act on it without opening a spreadsheet. On its own an objective cannot be judged, so it is paired with a measure: key results in OKRs, a KPI with a target in hoshin kanri and strategy deployment. A good objective names one direction, has one owner and says why it matters, and each level of the plant writes its own version of the level above rather than copying the words. Too many objectives is the commonest failure; three to five for a year is a usual ceiling.
OKR stands for objectives and key results, a goal-setting format in which a team writes one objective, the change it wants to make in words, and three to five key results, the measurable outcomes that will show the objective has been reached, then grades them at the end of a cycle, usually a quarter. The format comes from Intel in the 1970s and was spread by John Doerr, whose What Matters site describes OKRs as measures for change where KPIs are measures of health, written top down and bottom up so that every level links to the top priorities. On the shop floor an OKR suits the change work: a line team's objective might be to make changeovers predictable, with key results on changeover minutes and first-pass yield after restart, while the KPIs keep tracking the day to day. OKRs and hoshin kanri can run together: hoshin sets the three to five year breakthroughs and the year's objectives, and OKRs give a team a quarterly format for its share.
OKR scoring is the grading of each key result at the end of its cycle, usually a quarter, so the team can say plainly how far it got and what it learned. What Matters lists three common methods: a yes or no on each key result, a red, yellow or green rating, or Google's scale from 0.0 to 1.0, where 1.0 means the target was fully reached and the objective's score is the average of its key results. The score opens the review, it is not a performance rating: the owner says what the number means, what got in the way and what changes next cycle. Plants that run hoshin kanri instead judge each goal Met or Missed against its KPI every month, a stricter rule that gives an earlier signal than one score a quarter.
On-time delivery (OTD) is the share of deliveries that arrive within the agreed window: on-time deliveries divided by all deliveries, times 100. The window has to be agreed first, for example up to two days early and none late, and some companies count order lines instead of deliveries. It is used to rate suppliers and as a plant’s own delivery measure to its customers; on-time in full (OTIF) also requires the full quantity.
One-piece flow means moving products through a process one unit at a time, with each step making only what the next step needs and passing it on immediately, instead of in batches. It cuts work-in-process and lead time and exposes defects quickly, because a problem is found at the next step instead of after a whole batch is made. It usually needs a cell layout, balanced work, and reliable equipment.
A one-point lesson (OPL) is a single-page visual training sheet that teaches one piece of knowledge or one skill, such as how to read a gauge or set a guide. It is mostly pictures with little text, can be read in a few minutes at the machine, and is often written by operators or technicians themselves. OPLs are widely used in TPM and autonomous maintenance to share know-how quickly.
Operational excellence is the ability of an organization to deliver consistently on safety, quality, delivery, and cost while steadily improving, because good performance is built into how daily work is run. In manufacturing it combines lean, quality, and maintenance methods with standard work, daily management, and leadership routines that keep them going. It describes a way of operating rather than a single program or tool.
OSHA recordkeeping is the system 29 CFR Part 1904 requires most US employers with more than 10 employees to keep for work-related injuries and illnesses. Each recordable case goes on the 300 Log within seven calendar days, with a 301 incident report. After the year ends a company executive certifies the 300A summary, which is posted from February 1 to April 30, and the records are kept for five years. Some establishments also send their data to OSHA online by March 2.
Overall equipment effectiveness (OEE) measures how much of a machine’s planned production time is truly productive. It is calculated as availability × performance × quality: the share of planned time the machine runs, how fast it runs against its ideal rate, and the share of parts that are good. A score of 100% means only good parts, made as fast as possible, with no stop time.
Overall labor effectiveness (OLE) shows how well a crew's paid hours turn into good output. It multiplies availability (paid hours not lost to waiting or idle time), performance (output against the standard time per unit) and quality (good units out of all units made). It is the labor counterpart of OEE, and unlike OEE it has no single standard definition, so write down how you count idle time and where the standard comes from before you compare periods.
An overhead crane inspection is the scheduled check of a bridge or gantry crane: its controls, brakes and limit switches, hook, rope or chain, and the structure and parts it runs on. In the United States, OSHA 29 CFR 1910.179 divides inspections of cranes in regular service into frequent inspections at daily to monthly intervals and periodic inspections at 1 to 12 month intervals, with signed monthly records for hooks, hoist chains and running ropes. A deficiency that is a safety hazard is corrected before the crane is used again.
The P-F interval is the time from the point P, where a coming failure can first be detected (rising vibration, heat, a crack, metal in the oil), to the point F, where the equipment fails. It sets how often a condition check must run: no more than half the P-F interval, and shorter if the time left after a late detection is less than the time needed to plan and do the repair. A bearing whose vibration shows 6 weeks before failure is checked every 3 weeks.
The pacemaker process is the one process in a value stream that receives the customer schedule. Processes upstream of it are replenished by pull from supermarkets, and material downstream of it moves in continuous flow or a FIFO lane, so its pace sets the pace of the whole stream. It is usually the last step of continuous flow before the customer, and the place where the product mix is leveled.
Paperless manufacturing is running and reporting the daily work of a plant without paper forms, checklists and documents: checks, logs, audits and instructions are filled in and read on phones, tablets or screens. The data is entered once where the work happens, so the people who act on it do not wait for paper to be collected, typed up or filed. A common way to start is form by form, beginning with the forms that take the most time or wait longest before anyone reads them.
Pareto analysis ranks problems or causes by their impact, usually as a bar chart sorted from largest to smallest with a cumulative percentage line. It rests on the 80/20 rule of thumb that a few causes account for most of the effect. Teams use it to decide which defects, downtime reasons, or losses to work on first.
In quality, parts per million (PPM) is the number of defective parts for every million made, shipped or received: defective parts divided by total parts, times 1,000,000. Eight rejects in 10,000 parts received is 800 PPM. It is the usual way to state supplier and customer quality, because percentages become hard to read once defect rates are small.
The payback period is how long a project takes to earn back its investment: the investment divided by the annual net benefit, which is the yearly saving or extra income minus any yearly running costs. A $120,000 machine that returns $40,000 a year net pays back in 3 years. It is quick to work out and easy to explain, but it ignores the time value of money and everything after the payback date, which is why capital requests usually show NPV as well.
PDCA, short for Plan-Do-Check-Act, is the four-step cycle behind continuous improvement. You plan a change based on what you expect it to do, try it on a small scale, check the results against the expectation, and act by standardizing what worked or starting another cycle. It is also known as the Deming or Shewhart cycle, and some practitioners use PDSA, with Study in place of Check.
The four steps, round and round
1PlanA change, based on what you expect it to do
2DoTry it on a small scale
3CheckThe results against the expectation
4ActStandardize what worked, or start another cycle
Then another cycle
Plan a change based on what you expect it to do; Do, try it on a small scale; Check the results against the expectation; Act, standardize what worked or start another cycle.
People, process, technology (PPT) is a framework for planning and checking a change by looking at three elements together: the people who do the work, the processes that set how it is done and the technology that supports them. The point is balance: a new tool adds little if the people are not trained or the process was not changed to fit it, so the weakest element limits the result. It is also called the golden triangle.
Pitch is the time a process needs to make one container of product: takt time × pack-out quantity. With a takt time of 120 seconds and 15 pieces to a container, the pitch is 30 minutes. Pitch sets the interval for releasing and collecting work, for example how often a kanban is taken from a heijunka box.
A plan for every part (PFEP) is a table with one row per part number that records everything about how the part is handled: supplier, pack quantity, container size and weight, daily usage, lead time, where it is stored, where it is used and how it is replenished. The Lean Enterprise Institute calls it a critical tool of the Toyota Production System. Market sizes, delivery routes and kanban card counts are worked out from it, so it has to be kept up to date.
Planned downtime is time a machine or line is deliberately stopped on a schedule, for example for planned maintenance, changeovers, cleaning, trials, or breaks. Because it is known in advance, parts, people, and the production plan can be prepared for it. Plants set their own rules for which planned stops count against OEE, so the rule should be written down and applied consistently.
Planned maintenance is maintenance work that is prepared in advance: the job is defined, and the steps, parts, tools, skills, and time it needs are worked out before it starts. It covers preventive and predictive tasks as well as repairs that can be prepared ahead of time. In TPM, planned maintenance is also the pillar through which the maintenance team builds a proactive program.
Planned maintenance percentage (PMP) is the share of maintenance hours spent on work that was planned before it was done: planned hours ÷ total maintenance hours. Planned work includes preventive and predictive tasks, inspections and repairs scheduled ahead; the rest is reactive work after a breakdown. Plants draw the line between the two differently, so the definition should be written down next to the number.
PM compliance is the share of preventive maintenance tasks due in a period that were done on time: PMs done on time ÷ PMs due. Many plants count a PM as on time if it is done inside a grace window they set, such as 10% of the interval, and count late PMs as done but not on time. A falling PM compliance is often the first sign that breakdown work is crowding out planned work.
A points-based reward system gives employees a set number of points for each action it recognizes and lets them save the points and redeem them for rewards they choose from a catalogue. The workplace sets the points per action and per prize, so the rules are visible and small, frequent recognition adds up. Points earned and not yet redeemed are a future cost to budget for.
Poka-yoke, Japanese for mistake-proofing, is any device or method that prevents an error from being made, or makes it obvious at once so it cannot become a defect. Examples include fixtures that only accept a part the right way round, sensors that stop a machine when a step is skipped, and connectors that cannot be mated incorrectly. It builds quality in at the source instead of relying on inspection.
Poka-yoke devices are commonly grouped into three types by how they catch an error. Contact methods check a physical feature of the part, such as its shape, size, or position; fixed-value methods check that the right number of parts or actions has been completed; and motion-step methods check that the steps of a process were done in the right order. Each type can either prevent the error or warn when it happens.
A powered industrial truck is a fork truck, tractor, platform lift truck, motorized hand truck or other specialized industrial truck driven by an electric motor or an internal combustion engine, the scope of OSHA 29 CFR 1910.178. OSHA's eTool groups them into seven classes, from electric rider trucks (class I) to rough terrain forklifts (class VII). In the US each operator must be trained and evaluated in the workplace, with an evaluation at least once every three years.
A pre-shipment inspection is the last check of an order before it leaves the building: the right part and revision, the right quantity, released lots, labels in the customer’s format, packaging as specified, complete paperwork and photos, followed by a release by an authorized person. Final inspection checks that the product meets its specification; pre-shipment inspection checks that the right product is packed, labeled and documented correctly. It catches the complaints that are not about the part itself, such as wrong quantities, mixed lots and missing certificates.
Predictive maintenance (PdM) schedules maintenance based on the measured condition of the equipment rather than on the calendar. Readings such as vibration, temperature, oil analysis, or current draw are monitored for signs of deterioration, and the repair is planned before failure is expected. It avoids both unexpected breakdowns and replacing parts that still have useful life.
The prevention-appraisal-failure (PAF) model sorts the cost of quality into four categories: prevention costs such as training and error-proofing, appraisal costs such as inspection and testing, internal failure costs such as scrap and rework, and external failure costs such as returns and warranty claims. The two failure categories make up the cost of poor quality. Spending more on prevention usually reduces the other three.
Preventive maintenance (PM) is maintenance done at set intervals of time or usage, such as monthly or every set number of running hours, to keep equipment from failing. Typical tasks are inspections, lubrication, adjustments, and replacing wear parts before they fail. It reduces breakdowns but can replace parts earlier than needed, which is why it is often combined with predictive maintenance.
A prioritization matrix is a grid for ranking improvement ideas or problems, most often by impact on one axis and effort on the other. High-impact, low-effort items are done first, high-impact, high-effort items are planned as projects, and low-impact items are simplified or dropped. It turns a long backlog into an agreed order of work.
Process capability measures how well a stable process can produce output within its specification limits. Cp compares the width of the tolerance with the spread of the process, (USL − LSL) ÷ 6σ, while Cpk also accounts for centering and is the smaller of (USL − mean) ÷ 3σ and (mean − LSL) ÷ 3σ. A Cpk of 1.33 is a commonly used minimum (a rule of thumb, not a universal standard), and the result is only meaningful when the process is in statistical control.
A process capacity sheet lists each machine in a process with its manual time, automatic time and tool change time, and works out how many pieces it can make in a shift: available time divided by the time per piece, including each piece's share of the tool changes. It shows which machine limits capacity before the operator's work is designed around it with a combination table.
Process cycle efficiency (PCE) is the share of a process's lead time spent on value-added work: value-added time ÷ lead time, in the same units. For example, a part with 110 seconds of value-added work and a 17-day lead time has a PCE of about 0.01%. Very low figures are normal in batch production; they show that almost all of the lead time is waiting, which is where improvement should start.
A process flow diagram is a drawing of every step a product goes through, in order, from the material coming in to the product going out, using symbols for operations, moves, inspections, delays and storage, with the decisions and rework loops between them. Counting the steps of each kind shows how few of them change the part, and the process FMEA and control plan follow the same steps; in automotive PPAP it is a required element. Unlike a value stream map, it does not show cycle times, inventory between steps or information flow.
Process improvement, also called business process improvement (BPI), is the ongoing work of finding, analyzing and improving existing processes so they take less time and go wrong less often. Common methodologies include lean, kaizen, PDCA, 5 Whys, total quality management and Six Sigma, and the one that fits depends on the problem: its size, its cause and the data available.
Pp and Ppk are process performance indices: they use the same formulas as Cp and Cpk, but with the overall standard deviation of every reading, which includes shifts and drift between subgroups, instead of the short-term variation within subgroups. Pp = (USL − LSL) ÷ 6σ, and Ppk is the smaller of (USL − mean) and (mean − LSL), divided by 3σ. When a process is stable Ppk and Cpk are close; a Ppk well below Cpk means the process moves over time, and PPAP initial process studies usually report Ppk.
The production part approval process (PPAP) is how a supplier, mainly in automotive, shows its customer that it understands every requirement for a part and that its production process can make the part to those requirements at the quoted rate. The AIAG PPAP manual lists 18 elements, from the design records, process flow diagram and FMEAs to capability studies and the part submission warrant (PSW), and five submission levels that set what is sent to the customer and what is kept on file. A new submission is usually needed for a new part and for changes such as new tooling, a new material or a move to another site.
A project charter is the one-page agreement that starts an improvement project, signed by the sponsor before the analysis begins. It states the business case, the problem in numbers, the goal with a date, the metrics, the scope, the team and the phase milestones. In DMAIC it is the main output of the Define phase and is updated at a tollgate if the data changes the picture.
A Pugh matrix compares several concepts with one reference concept, the datum, one criterion at a time, marking each as better (+), the same (S) or worse (-). The team counts the marks, studies the weaknesses, improves or combines the concepts and repeats with a stronger datum. It is named for Stuart Pugh and is also called controlled convergence.
A pull system is a way of controlling production in which a process makes or moves material only when the next process signals that it has used some, instead of following a forecast-driven schedule (push). The signal can be a kanban card, an empty container or space, or an electronic trigger. Because work is released only as it is consumed, inventory stays capped and overproduction is avoided.
A pulse survey is a short survey, often a dozen or so statements, run at regular intervals to see how people view their work between longer surveys. Because it is short and repeated with the same statements, results can be compared over time and actions checked against the next round.
A QA matrix (quality assurance matrix) is a table of defect modes against the process stations in flow order that marks where each defect is made and where it is first caught. It shows which stations make defects, which checks catch them, how far a defect travels before it is found and which defects reach the customer, and ratings rank the defects so the most important are fixed first. It is common in TPM and World Class Manufacturing quality work, and each plant sets its own format and scales.
A quality alert is a notice posted at the line after a defect is found, showing what the defect looks like next to a good part, how to check for it and what to do if it is found. Every operator on every shift is briefed and signs it, and it stays up until its review date or until the cause is fixed. It is a containment step, not a fix: the cause is dealt with through an NCR or a CAPA.
Quality at the source means each person and process makes sure their own output is right before passing it on, instead of relying on inspection at the end. Operators are given the standards, gauges, and error-proofing to check their own work, and they can stop the process when something is wrong. It is the principle behind jidoka and poka-yoke.
A quality circle is a small group of employees from the same work area, usually with their supervisor, who meet regularly to find, analyze, and solve problems in their own work. Members are trained in simple problem-solving tools such as Pareto charts, fishbone diagrams, and check sheets, and they carry out fixes within their area of authority. The practice is associated with Kaoru Ishikawa and spread widely through Japanese industry.
A RACI matrix is a table that lists the tasks or decisions in a process against the people or roles involved and marks each as responsible (does the work), accountable (owns the outcome and signs off), consulted (gives input beforehand), or informed (kept up to date). Each task should have exactly one accountable person. It removes confusion about who does what, especially in cross-functional projects and handoffs.
Reactive maintenance means repairing equipment only after it has broken down, also called run-to-failure or breakdown maintenance. It can be a sensible choice for cheap, non-critical items, but as the main strategy it leads to unplanned downtime, rushed repairs, and higher costs. Proactive approaches, such as preventive, predictive, and autonomous maintenance, aim to catch problems before failure.
REBA, the Rapid Entire Body Assessment, is an observational method published by Sue Hignett and Lynn McAtamney in 2000 for scoring the risk of one whole-body working posture. The assessor scores the trunk, neck, legs, upper arm, lower arm and wrist, adds the load, the hand coupling and an activity score, and reads a score from 1 to 15 with five action levels. It screens postures; it does not measure how long or how often they are held.
Red tagging is the standard technique for the Sort step of 5S. Any item whose need in the area is in doubt gets a red tag and is moved to a holding area, where it stays for a set period before being returned, relocated, or disposed of. It clears out unneeded tools, materials, and equipment without people having to decide everything on the spot.
Refresher training is training repeated to keep a qualification current, either at a set interval or after a trigger such as a change of job, machine or procedure, an incident, or a failed check. Some rules set the interval, for example an evaluation of each forklift operator at least once every three years under OSHA 29 CFR 1910.178(l)(4)(iii); many intervals are set by the site, the customer or the training provider. A training matrix with expiry dates shows when each one is due.
Reliability-centered maintenance (RCM) is a structured method for choosing the right maintenance approach for each asset based on its functions, how it can fail, and the consequences of each failure. For every failure mode the team picks a task, such as condition monitoring, scheduled replacement, a redesign, or deliberately running to failure, that is technically feasible and worth doing. It focuses maintenance effort on the failures that matter most for safety, production, and cost.
Rework is extra work done on a unit that failed its check so it can be made good, such as a second wash, a hand deburr or a re-press. It costs the extra minutes times the labor rate, not the whole unit, which is why it is counted apart from scrap. A reworked unit still counts against first pass yield, so rework is how a good final yield can hide a poor process.
Right first time (RFT) is the goal, and the measure, of doing every task correctly the first time, with no rework, repair, correction, or repeat. As a metric it is the share of units, batches, or documents that pass without any correction, much like first pass yield. It is achieved by building quality into the process through capable processes, clear standards, and error-proofing rather than by inspection.
Rolled throughput yield (RTY) is the probability that a unit passes through every step of a multi-step process without any rework or scrap. It is calculated by multiplying the first pass yields of the steps, so three steps at 95% FPY each give an RTY of about 86%. Because losses compound, RTY shows the end-to-end quality of a line better than any single step’s yield.
Root cause analysis (RCA) is the process of finding the underlying reason a problem occurred, so the fix prevents it from happening again instead of treating the symptom. Common tools include the 5 Whys, fishbone diagrams, Pareto analysis, and fault tree analysis. A root cause is confirmed when removing it stops the problem from coming back.
A safety cross is a calendar in the shape of a cross, with one square for each day of the month, that a team colours each day by the most serious safety event of that day. The colours are set by the site, for example green for no incident and red for an injury, and the cross is reviewed at the daily huddle so the month's record and any pattern show at a glance. A green day only means nothing was reported, so it is read together with the near miss count.
Safety stock is the extra inventory held to cover variation in demand and in supplier lead time, so a stockout does not happen between placing an order and receiving it. A common formula multiplies a z-score for the service level wanted by the standard deviation of demand over the lead time. The reorder point is the average demand during the lead time plus the safety stock.
Sangen shugi, Japanese for the “three actuals” principle, is the habit of checking the actual place (genba), the actual thing (genbutsu) and the actual facts (genjitsu) before deciding anything about a problem. It is genchi genbutsu written as a checklist. The 5G methodology adds two more: the principles that govern the process (genri) and the standard it should follow (gensoku).
A scatter diagram plots pairs of measurements as points on two axes to show whether one variable changes with another, for example oven temperature against the number of defects. Points that cluster along a line or curve suggest a relationship, while a shapeless cloud suggests none. A pattern shows correlation, not cause, so a suspected cause still has to be confirmed with a test or at the gemba.
Maintenance schedule compliance is the share of the jobs on the week's schedule that were done as scheduled: jobs done as scheduled ÷ jobs scheduled, or the same with the jobs' estimated hours. 14 of 16 scheduled jobs is 87.5%. It is not the same as PM compliance, which counts only preventive tasks against their due dates, and sites differ on whether a job counts if it was done in its week or only on its day.
Scheduled maintenance decides when planned maintenance work will be done and who will do it, fitting jobs into production windows and technician availability. Planning answers what and how; scheduling answers when and by whom. The two terms are often used interchangeably, but they are separate steps, and a job should be planned before it is scheduled.
Scrap rate is the share of the units made that end up as scrap: scrapped units divided by the total units made, good and bad, times 100. Count only what is thrown away, and track reworked units separately, because rework costs time but not the whole unit. The cost of scrap is the scrapped units times what each one already cost in material, labor and machine time.
Scrum is a framework in which a small team works in sprints of one month or less, each with one sprint goal, and holds four timed events in every sprint: sprint planning, the daily scrum, the sprint review and the sprint retrospective. The roles are the Product Owner, the Scrum Master and the people doing the work, called developers. It is set out in the Scrum Guide by Ken Schwaber and Jeff Sutherland, last revised in 2020.
The seven basic quality tools are a set of simple charts and methods for collecting and analyzing quality data: the check sheet, Pareto chart, cause-and-effect (fishbone) diagram, histogram, scatter diagram, control chart, and stratification. Kaoru Ishikawa promoted them in Japan so that front-line teams, not only specialists, could solve most everyday quality problems themselves. Some lists replace stratification with a flowchart.
A shadow board is a board with the outline of each tool drawn or cut out where it hangs, so every tool has a fixed home and a missing one is noticed at once. It is a common Set in order tool in 5S and a simple example of visual management. Boards are kept at the point of use, and the outlines are often labeled with the tool name or size.
Single-minute exchange of die (SMED) is a method for cutting changeover time, the time from the last good part of one product to the first good part of the next. Its core idea is to separate internal work, which can only be done with the machine stopped, from external work that can be done while it runs, then move as much as possible to external and streamline both. The name reflects the goal of changeovers under ten minutes, a single-digit number of minutes.
What SMED shrinks: the machine-down window
Machine-down time at each stage. Baseline: machine down 18 blocks, running alongside 0; After step 3: machine down 9 blocks, running alongside 6; After step 4: machine down 6 blocks, running alongside 4; Standard: machine down 6 blocks, running alongside 4.
18
0
Baseline
9
6
After step 3
6
4
After step 4
6
4
Standard
Machine downMachine running
Internal work, machine stopped
External work, machine running
Blocks, not minutes: illustrative.
Only the part above the line is changeover time; the work below it happens while the machine runs.
Four stages. Machine down, above the line: 18 blocks at the baseline, 9 after external work moves out of the stop, 6 after the remaining work is shortened, 6 held as the standard. Machine running, below the line: the external work done alongside production, 6 then 4 blocks. Blocks are illustrative, not minutes.
A SIPOC diagram is a one-page, high-level map of a process that lists its suppliers, inputs, process steps, outputs, and customers in five columns. The process column is kept to a handful of main steps, so the team agrees on where the process starts and ends, and who it serves, before mapping it in detail. It is commonly used in the Define phase of DMAIC and at the start of improvement projects.
The six big losses are the main causes of lost productivity on equipment, grouped under the three OEE factors: equipment failures and setup and adjustments (availability), idling and minor stops and reduced speed (performance), and process defects and reduced yield at startup (quality). Naming losses this way tells a team where OEE is being lost and what kind of fix is needed.
Six Sigma is a data-driven method for reducing variation and defects in a process, organized around the DMAIC cycle and statistical tools. The name refers to a process so well controlled that it produces no more than 3.4 defects per million opportunities. Lean Six Sigma combines it with lean’s focus on flow and waste, and practitioners are trained in levels called belts, such as Green Belt and Black Belt.
A skills matrix, also called a competency or training matrix, is a grid that shows each employee’s level of proficiency in each skill or task their area needs. It shows at a glance who is trained, who is in training, and where the team is thin, so leaders can plan training and cross-training. Most matrices use a simple scale, for example from not trained to able to train others.
SMART goals are goals written to pass five tests: Specific, Measurable, Achievable, Relevant and Time-bound. In a plant that means one metric with a baseline and a target, a named line or area and an owner, a link to a plant objective and a deadline, for example cutting scrap on Line 3 from 5.0% to 3.0% of parts made by 31 December. Progress is then reviewed against the target on dates set in advance.
A spaghetti diagram is a drawing of the actual path a person, product, or document takes through a work area, traced as one continuous line on a floor layout. The tangle of lines makes wasted walking, transport, backtracking, and crossing flows visible, and the total distance can be measured. Teams draw one at the gemba before and after a layout change to show what it saved.
SQDCM stands for safety, quality, delivery, cost, and morale, the five categories many plants use to organize daily performance boards and tiered meetings. Each category has a few simple measures that are reviewed every shift or day against target, and misses are turned into actions. Some sites use SQDC without morale, or swap in people or environment.
A standard maintenance procedure (SMP) is a documented, step-by-step method for doing a specific maintenance task, such as replacing a bearing or calibrating a sensor, the same safe way every time. It usually lists the tools, parts, safety precautions including lockout points, and the checks that confirm the job was done right. SMPs turn experienced technicians’ know-how into a method others can learn and repeat.
A standard operating procedure (SOP) is a written, approved document that describes how to carry out a routine operation consistently and safely. It typically states the purpose, scope, responsibilities, required materials, and the steps in order, and it is kept under document control so only the current version is used. On the shop floor, SOPs are often backed by shorter work instructions at each station.
Standard work is the current best known way to do a job, written down with three elements: takt time, the work sequence, and the standard in-process inventory needed to run it. It gives a stable baseline, so abnormalities stand out and improvements can be measured and locked in. Standard work is expected to change as the team finds better methods.
A standard work audit is a periodic check that the written standard for a job matches what happens at the station and is still the best known method. The auditor compares the posted document, the work sequence, cycle times, training records and tools with the standard, then decides for each gap whether the operator drifted from a good standard or the standard itself needs updating, because the two call for opposite fixes. It complements the daily checks in leader standard work.
A standard work combination table is a chart of one operator's work elements in sequence, with the manual, machine and walking time of each drawn on a time scale against takt. It shows how the operator's work fits with the machines' automatic cycles, where the operator waits, and whether the cycle fits within takt. It is one of the core standard work documents, with the process capacity sheet and the standard work chart.
Standard work in process (SWIP) is the minimum number of parts that must be inside a process, at and between stations, for operators to follow their standard work sequence without waiting. It is one of the three elements of standard work, with takt time and the work sequence. A common estimate is the total cycle time of the steps ÷ takt time, plus any parts held for curing, cooling, or inspection; anything above the standard is excess inventory.
Statistical process control (SPC) is the use of statistics, mainly control charts, to monitor a process while it runs and tell normal variation from real change. A control chart plots measurements over time against a center line and control limits, usually set three standard deviations either side of the mean. Points outside the limits or non-random patterns signal a special cause to investigate, while variation inside them is common cause and can only be reduced by changing the process.
Stop-work authority is a written company rule that any person on site, employee or contractor, may stop a task, their own or someone else's, when they believe someone could be hurt, and that nobody is punished for a stop made in good faith, even one that turns out to have been unnecessary. Work restarts only when a named person in charge has confirmed the hazard is gone or controlled and the crew agrees, and the decision is recorded. It is wider than the legal right to refuse dangerous work, which in the United States protects a worker only where a reasonable person would see a real danger of death or serious injury. In US offshore oil and gas it is required by regulation (30 CFR 250.1930); elsewhere it is policy. It is the safety counterpart of the andon cord.
Strategy deployment is the work of turning a company's strategy into a few goals, giving each a measure and an owner, carrying them down to every facility, department and person, and reviewing them on a fixed rhythm until they are met. Hoshin kanri is the lean method for it, with the X-matrix, catchball and the bowling chart; OKRs and balanced scorecards are other formats, and a plant may combine them. The point is that a team leader can show how today's board connects to the company's goals, and leadership can see which goals each site is meeting and where to help. Sull, Sull and Yoder found in MIT Sloan Management Review (2018) that only 28% of the executives and middle managers responsible for executing strategy could list three of their company's strategic priorities, the gap that deployment exists to close.
A strategy map is a one-page diagram of a strategy in which the objectives sit in layers and arrows show which objective is expected to cause which. Robert Kaplan and David Norton described it in the Harvard Business Review in 2000 as an extension of their balanced scorecard, with four layers: financial and customer for the results wanted, internal process and learning and growth for how to reach them. Each objective carries a measure, a target and an owner. In a plant it ties a skill such as cross-training on changeovers to a schedule, a delivery result and a cost, and it is the reasoning that an X-matrix then turns into a plan with owners.
Stratification is the practice of splitting data into groups by its source, such as machine, shift, operator, supplier, or product, before analyzing it. A problem that looks spread evenly across a whole process often turns out to come mostly from one group once the data is separated. It is one of the seven basic quality tools and is usually applied to check sheet, Pareto, and scatter diagram data.
The substitution test asks whether a different person doing the same job, with comparable experience and qualifications, would have acted the same way given how events looked at the time. If the honest answer is probably yes, blame has no useful role and the finding is about the task, the training or the equipment; if the answer is no, the next question is whether something in the system produced the difference before negligence is considered. It was proposed by Neil Johnston in 1995 and built into James Reason's culpability decision tree, and it is meant to be answered by peers, one at a time, and written down.
In lean, a supermarket is a controlled store of parts placed between two processes that cannot be linked by continuous flow. The downstream process takes what it needs, and the upstream process makes only enough to replace what was taken, usually signaled by kanban. Each item has a fixed location and set minimum and maximum quantities, so the stock level is visible and capped.
A supplier audit is a visit to a supplier’s site to check that the quality system and the process that makes your parts are in place and working, scored against a checklist. A system audit checks the quality management system against a standard such as ISO 9001; a process audit follows a part from receiving through production, inspection and shipping. Audits are done before approving a new supplier, after serious quality problems and on a schedule set by risk, and the findings come back as an action plan with dates.
A supplier corrective action request (SCAR) is a formal request to a supplier to find and fix the root cause of a problem with what they supplied, such as a rejected lot or a failure traced to their part. It describes the nonconformance and asks for containment, root cause, corrective action and proof that the action worked, often in an 8D format, by set dates. Days to respond and the number of open SCARs are common measures on a supplier scorecard.
A supplier scorecard is a regular, usually monthly, report card that rates each supplier on the same measures, most often quality (PPM or lot rejection rate), on-time delivery and how quickly corrective action requests are answered. Each measure is scored against a target, weighted and added into one score and rating, so supplier reviews start from data both sides can check. The trend shows which suppliers need development and whether their corrective actions are working.
The Swiss cheese model is James Reason's picture of how accidents happen in organisations: the defences against a hazard are slices with holes that open and move, and harm gets through when holes in several slices line up. The holes are active failures, the unsafe acts of people at the sharp end, and latent conditions, weaknesses such as poor design, a maintenance backlog or an unworkable procedure that sit in the system for months. It directs an investigation past the last act to the conditions that let it through.
Takt time is the rate at which products must be finished to meet customer demand, calculated as available production time divided by customer demand for that period. For example, 27,000 seconds of available time and demand for 900 units give a takt time of 30 seconds per unit. It sets the pace for the line and is the target that station cycle times are balanced against.
The example, worked
Takt time = available production time ÷ customer demand
27,000 s ÷ 900 units
Takt time30 s per unit
Takt time equals available production time divided by customer demand: 27,000 seconds divided by 900 units is 30 seconds per unit.
A target condition is a description of how a process should work at a set date not far off, with the measurable result it should reach. In the improvement kata it sits between the current condition and the longer-term challenge, and the learner experiments toward it one obstacle at a time. It describes the work pattern that should produce the number, not only the number itself.
A team charter is a short written agreement, made by a standing team with its sponsor, that sets out the team's purpose, scope, members and roles, the KPIs it owns, the decisions it may make on its own, its norms and its meeting cadence. Unlike a project charter it has no end date: the team reviews it on a set date and whenever its scope or members change.
The theory of constraints (TOC) is a management approach, developed by Eliyahu Goldratt, that treats the output of any system as limited by its tightest constraint, often a bottleneck machine or process. Improvement follows five focusing steps: identify the constraint, exploit it, subordinate everything else to it, elevate it, and then repeat for the next constraint. Time lost at the constraint is lost for the whole system, so that is where improvement pays back most.
Throughput is the rate at which a process turns out finished, good output, such as units per hour or orders per day. A line’s throughput is set by its bottleneck, so speeding up other steps does not raise it and only builds inventory. In the theory of constraints, throughput is defined in money terms instead, as the rate at which the system generates money through sales.
Throughput time is the total time a unit takes to pass through a process from start to finish, including processing, inspection, moving, and waiting. Because waiting is usually the largest part, it is much longer than the sum of the cycle times, and cutting queues and batch sizes shortens it most. Despite the similar name it is a time, not a rate like throughput, and for a stable process it equals work in process ÷ throughput (Little’s Law).
A tier board is the visual board a team stands at during its tiered meeting, showing that level’s performance against target, usually grouped by SQDCM, along with open problems and their actions. Each tier’s board is fed by the one below it, so an issue can be traced from the front-line board up to the plant leadership board. Red and green marks against each target show in seconds where attention is needed.
Tiered meetings are a cascade of short stand-up meetings held at set times each day, from front-line teams (tier 1) to area leaders (tier 2) and plant leadership (tier 3 and above). Each tier reviews performance, often by SQDCM, and escalates problems it cannot solve to the next tier, while decisions and support flow back down. The cascade gets each problem to the level that can fix it on the same day.
The cascade of tiers
Plant leadershipTier 3 and aboveDecisions and support flow back down
Area leadersTier 2Escalates what it cannot solve
Front-line teamsTier 1Short stand-up at a set time each day
Problems go up a tier; decisions come back down, the same day.
Tier 3 and above, plant leadership; tier 2, area leaders; tier 1, front-line teams. Each tier meets every day at a set time and escalates problems it cannot solve to the next tier, while decisions and support flow back down.
A time study is a work measurement method: an observer times each element of a job over a number of cycles, rates the operator's pace against a defined standard and adds allowances, to get the standard time for a qualified worker. How many cycles to time depends on how much the readings vary and how close to the true average the result must be, for example within 5% at 95% confidence. The result feeds standard work, line balancing and the comparison with takt time.
Total effective equipment performance (TEEP) measures how much of all calendar time, 24 hours a day, every day of the year, a machine spends making good parts at full speed. It is calculated as OEE × utilization, where utilization is planned production time ÷ all available time. OEE shows how well the scheduled time is used, while TEEP also shows the capacity left in unscheduled shifts, weekends, and planned stops.
Total productive maintenance (TPM) is a plant-wide approach to equipment reliability that makes operators, maintenance, and management jointly responsible for keeping machines in top condition. It aims for no breakdowns, no small stops or slow running, no defects, and no accidents, and it is usually organized in pillars such as autonomous maintenance, planned maintenance, and focused improvement. OEE is its usual measure of progress.
Total quality management (TQM) is a management approach in which everyone in the organization shares responsibility for quality and for continually improving processes to meet customer needs. Its core principles include customer focus, involvement of all employees, a process view of work, decisions based on data, and continual improvement. Many of its tools, such as quality circles, Pareto analysis, and PDCA, are shared with lean and Six Sigma.
The total recordable incident rate is the number of recordable injuries and illnesses multiplied by 200,000 and divided by the hours worked by employees in the period. The 200,000 base is the hours of 100 full-time employees working 40 hours a week for 50 weeks, so the rate reads as cases per 100 full-time workers a year. Hours worked leave out paid time not worked, such as vacation, sick leave and holidays.
The Toyota Production System (TPS) is the production and management system developed at Toyota and the basis of what became known in the West as lean manufacturing. It is often drawn as a house with two pillars, just-in-time and jidoka, standing on a foundation of standardized work, heijunka, stability, and kaizen. Its aim is to make exactly what the customer needs with the least waste, while developing people who solve problems every day.
TWI Job Instruction is a four-step method for teaching someone to do a job correctly, safely, and consistently: prepare the worker, present the operation, try out performance, and follow up. Before training, the trainer writes a job breakdown sheet listing the important steps, the key points, and the reasons for each key point. It comes from the Training Within Industry program created in the United States during World War II and is widely used to teach standard work.
Unplanned downtime is any time a machine or line stops when it was scheduled to run, because of a breakdown, missing material or operators, a quality problem, or another unexpected event. It is the most disruptive kind of stoppage because nothing is ready for it, and in OEE it counts as an availability loss. Reducing it is the main goal of preventive, predictive, and autonomous maintenance.
An unsafe act is something a person does, or leaves undone, that raises the chance or the severity of harm: reaching past a guard, skipping lockout, working without the required PPE, taking a shortcut through a forklift lane. The term comes from H. W. Heinrich's 1931 claim that most accidents were caused by unsafe acts, a claim whose data cannot be checked and that modern safety practice has set aside. Most unsafe acts are the visible end of an unsafe condition and of the system behind it, so a report should record the act, the condition and what made the act sensible at the time.
An unsafe condition is something about the workplace or the equipment that could hurt someone whether or not anyone does anything wrong: an unguarded machine, a wet floor, a blocked exit, a missing lockout point, poor lighting. It is there when nobody is near it, which is the test that separates it from an unsafe act. Conditions can be designed out, guarded or removed, so they are fixed at the top of the hierarchy of controls.
Value stream mapping (VSM) is a lean method for drawing every process step, material flow, and information flow needed to bring a product from order to delivery. The current-state map shows where time is lost and waste sits, usually summarized as total lead time against value-added time; the future-state map shows the target flow and the plan to reach it. It helps teams improve the whole flow instead of optimizing single steps.
Value-added time is time spent on steps that change the product in a way the customer is willing to pay for, done right the first time. Non-value-added time is everything else, and it splits into pure waste, which can be removed, and necessary non-value-added work, such as some inspections or regulatory steps, which cannot be removed yet. Comparing value-added time with total lead time shows how much of a process is waiting and handling.
Visual management is the practice of making the state of work visible at a glance, so anyone can tell whether things are normal or abnormal without asking. It includes floor markings, shadow boards, labels, andon lights, and performance boards. The point is to prompt fast action on problems, not just to display information.
A waste walk is a planned walk through a work area with the single purpose of spotting and recording waste, usually using the eight wastes as a checklist. Participants observe the real process, note each example with its location and evidence, and turn what they find into improvement actions with owners. It differs from a general gemba walk by having one narrow focus: finding waste.
A weighted decision matrix compares options against the same criteria: each criterion gets a weight that reflects how much it matters, each option is scored on each criterion, and the weighted scores are added up to rank the options. Agreeing the criteria and weights before scoring keeps the result from being steered toward a favorite. Many teams add pass or fail must-haves that rule an option out whatever its total.
Work in process (WIP) is material that has entered a production process but is not yet finished, such as parts waiting between stations, sitting in queues, or partly assembled. Lean treats excess WIP as inventory waste, because it ties up cash and floor space, hides problems, and lengthens lead time. By Little’s Law, average lead time equals average WIP ÷ throughput, so cutting WIP at the same output shortens lead time.
In maintenance, a work order is the record that authorizes and describes a job: what needs doing, on which asset, by whom, by when, and with what parts and safety steps. It is raised from a breakdown, an inspection finding, or a preventive maintenance schedule, and it is closed with the work done, the time spent, and the parts used. Closed work orders build the repair history used to calculate MTBF and MTTR and to plan future maintenance.
World class manufacturing (WCM) is a structured production system that combines lean, TPM, and total quality methods and aims for zero accidents, zero defects, zero breakdowns, and zero waste. It is built on ten technical pillars: safety and health, cost deployment, focused improvement, autonomous activities, professional maintenance, quality control, logistics, early equipment management, people development, and environment. Its hallmark is cost deployment, which puts a money value on losses so improvement work goes where the cost is highest.
An X-matrix is a one-page hoshin kanri planning tool that links long-term breakthrough objectives, annual objectives, top-level improvement priorities, and the metrics used to track them, with the people responsible shown alongside. The four lists sit around a central X, and grids in the corners mark which items support which. It lets leaders check that every improvement project traces back to a strategic goal.
A yamazumi chart is a stacked bar chart of the work content at each operator or station, broken into individual tasks and often colored by value-added, non-value-added, and waste. Drawn against a line for takt time, it shows at a glance which stations are overloaded or underused. Yamazumi means “stack up” in Japanese, and the chart is the standard tool for line balancing.
Yokoten is the practice of sharing a proven improvement or lesson learned sideways, across lines, shifts, and sites, so other teams can adopt and adapt it. The Japanese word roughly means “horizontal deployment.” Unlike a top-down mandate, the receiving teams go and see the original improvement and then fit it to their own conditions.
Zero breakdowns is the TPM aim of eliminating equipment breakdowns with five countermeasures: maintain basic conditions (cleaning, lubrication, tightening), keep to operating conditions, restore deterioration, correct design weaknesses and improve operating and maintenance skills. Each countermeasure answers one cause of breakdowns, so a breakdown analysis that names the missing one points straight at the fix.
Every article and e-book we have published is listed in the library, or bring a problem you are working on and we will show you how LeanSuite handles it.
